3 ms·

Signed up, I think this is an interesting and long overdue idea. Let's see if this shot works. Truth is there are a ton of CRUD apps out there, that are 80% bo

Signed up, I think this is an interesting and long overdue idea. Let's see if this shot works. Truth is there are a ton of CRUD apps out there, that are 80% boilerplate in reality, it's just different boilerplate. I know HN tends to focus more on startups and bleeding edge hobbyists but enterprise CRUD is what drives software development, and creates most of our checks. And we can still do much better. Whoever can come up with an elegant, fast, stable solution to crack out CRUD apps and not make all the mistakes of the past attempts can revolutionize things. In my opinion the LAMP stack did this great for a time. Rails does a pretty good job now. The .Net framework also does this very well, at a cost. It's not sexy, hip or cutting edge but there's a big demand for it.

couchand 12y ago
You want to talk about a very unsexy way to build CRUD apps fast, go look at Salesforce. Boy, do I hate developing on their platform, but it is a pretty damn quick way to build CRUD apps, as long as you're fine paying $100 per user per month.
- icebraining 12y ago
  And if you want something open source which you can install on your own machines, Odoo[1] is a nice alternative (disclaimer: I work for a partner company). It works in a similar way; you just define the model with its fields (as a Python class), add a view where you just indicate which fields to expose and where, and then a menu entry to point to that model, and done! You can even skip the view if you don't mind it generating a basic one for you. [1] https://www.odoo.com/ https://www.odoo.com/
  - couchand 12y ago
    That is a remarkable suite of tools available, and all for free. I'm surprised I haven't heard of it before, but I'll certainly be looking into it. Thanks for the tip!
delluminatus 12y ago
Absolutely. I even think this, or something like it, will catch on in enterprise development shops for those fringe projects that are done in like a week. Especially if it can be deployed on the same tech stack as that other random thing we built. We use ASP.NET MVC for a lot of this stuff, which combined with ADO.NET and SignalR and a good JS framework, gets you 75% of the way there. I will tell you what I would like it to be. Maybe it gives you a Visual Studio-style editor for websites, which let you easily lay out a page, edit CSS, add event handlers to elements, link to other pages, and make and use data bindings (which could be configurable: persistent to DB, associated with user account, validated with this function, etc.) Then you could use the data model created by the bindings, modify it if you want (data model specification is probably integral to everything), add and configure whatever extra stuff is supported (accounts, websockets, whatever). A cross between Weebly and Visual Studio. It would be great for demo apps and internal utilities.
- gfmio 12y ago
  This is quite accurately describing what we're trying to do on the front-end side. :) In addition to just creating RESTful APIs and backend and persistence code, we allow you to design your UI in a GUI (or using code). Views are data-driven, so you can essentially bind data from your models using drag&drop (or in code). However, instead of just using XY positioning, we focus on relative positioning of subviews (representing lower level parts of your application) to create working, well-designed interfaces (again, based on your design settings/parameters).
- vertex-four 12y ago
  Visual Studio Lightswitch is a thing, included in Visual Studio 2012 Professional onwards. It's essentially a tool for building a database (or interfacing to an external one) and putting a CRUD UI on top of it, while it allows for extending it with real code nearly everywhere. It also integrates into other enterprise-y stuff, supposedly. The issue is... you're stuck with Lightswitch once you start using it. There's not an "export to standalone C# project" button, I believe; you'll always be dependent on the Lightswitch libraries, which Microsoft will probably decide to drop or radically change in the future, at least before .Net itself is dropped. And that's going to be sort of the problem with all these systems. Unless it can output pure, maintainable C# code (or whatever else), not dependent on significant libraries specific to the RAD toolkit, there's a very high risk that the toolkit and libraries become unsupported well before the application stops being used.

couchand 12y ago

You want to talk about a very unsexy way to build CRUD apps fast, go look at Salesforce. Boy, do I hate developing on their platform, but it is a pretty damn quick way to build CRUD apps, as long as you're fine paying $100 per user per month.

icebraining 12y ago
And if you want something open source which you can install on your own machines, Odoo[1] is a nice alternative (disclaimer: I work for a partner company). It works in a similar way; you just define the model with its fields (as a Python class), add a view where you just indicate which fields to expose and where, and then a menu entry to point to that model, and done! You can even skip the view if you don't mind it generating a basic one for you. [1] https://www.odoo.com/ https://www.odoo.com/
- couchand 12y ago
  That is a remarkable suite of tools available, and all for free. I'm surprised I haven't heard of it before, but I'll certainly be looking into it. Thanks for the tip!

delluminatus 12y ago

Absolutely. I even think this, or something like it, will catch on in enterprise development shops for those fringe projects that are done in like a week. Especially if it can be deployed on the same tech stack as that other random thing we built. We use ASP.NET MVC for a lot of this stuff, which combined with ADO.NET and SignalR and a good JS framework, gets you 75% of the way there. I will tell you what I would like it to be. Maybe it gives you a Visual Studio-style editor for websites, which let you easily lay out a page, edit CSS, add event handlers to elements, link to other pages, and make and use data bindings (which could be configurable: persistent to DB, associated with user account, validated with this function, etc.) Then you could use the data model created by the bindings, modify it if you want (data model specification is probably integral to everything), add and configure whatever extra stuff is supported (accounts, websockets, whatever). A cross between Weebly and Visual Studio. It would be great for demo apps and internal utilities.

gfmio 12y ago
This is quite accurately describing what we're trying to do on the front-end side. :) In addition to just creating RESTful APIs and backend and persistence code, we allow you to design your UI in a GUI (or using code). Views are data-driven, so you can essentially bind data from your models using drag&drop (or in code). However, instead of just using XY positioning, we focus on relative positioning of subviews (representing lower level parts of your application) to create working, well-designed interfaces (again, based on your design settings/parameters).
vertex-four 12y ago
Visual Studio Lightswitch is a thing, included in Visual Studio 2012 Professional onwards. It's essentially a tool for building a database (or interfacing to an external one) and putting a CRUD UI on top of it, while it allows for extending it with real code nearly everywhere. It also integrates into other enterprise-y stuff, supposedly. The issue is... you're stuck with Lightswitch once you start using it. There's not an "export to standalone C# project" button, I believe; you'll always be dependent on the Lightswitch libraries, which Microsoft will probably decide to drop or radically change in the future, at least before .Net itself is dropped. And that's going to be sort of the problem with all these systems. Unless it can output pure, maintainable C# code (or whatever else), not dependent on significant libraries specific to the RAD toolkit, there's a very high risk that the toolkit and libraries become unsupported well before the application stops being used.